ORIGINAL PAPER

Distribution and community-level effects of the Chinesemystery snail (Bellamya chinensis) in northern Wisconsinlakes

Christopher T. Solomon Æ Julian D. Olden ÆPieter T. J. Johnson Æ Robert T. Dillon Jr. ÆM. Jake Vander Zanden

Received: 20 November 2008 / Accepted: 24 August 2009 / Published online: 10 September 2009

� Springer Science+Business Media B.V. 2009

Abstract Managing invasive species requires infor-

mation about their distributions and potential effects,

but community-level impacts of invasive animals

remain poorly understood. The Chinese mystery snail

(Bellamya chinensis) is a large invasive gastropod that

achieves high densities in waters across North Amer-

ica, yet little is known about its ecological significance

in invaded systems. We surveyed 44 lakes to describe

the patterns and determinants of B. chinensis distribu-

tions in northern Wisconsin, USA, and to assess the

likelihood of effects on native snail communities in the

invaded systems. B. chinensis was widespread among

surveyed lakes (21 of 42 lakes with snails) and its

occurrence was correlated with indicators of lake

productivity and anthropogenic dispersal vectors (boat

landings, distance to population centers, shoreline

housing density). Some native snail species tended not

to occur at sites where B. chinensis was abundant;

among these was Lymnaea stagnalis, which suffered

reduced survival in the presence of B. chinensis in a

recently published mesocosm study. However, there

was no difference in overall snail assemblage structure

at either the site or lake level as a function of

B. chinensis presence or abundance. Lake occurrences

of many snail species have apparently been lost over

time, but a comparison to a 1930s survey showed that

there was no increased likelihood of species loss in

lakes invaded by B. chinensis (or by the invasive

crayfish Orconectes rusticus). Although B. chinensis is

widespread and sometimes abundant in northern

Wisconsin lakes, it does not appear to have strong

systematic impacts on native snail assemblages.

Keywords Gastropoda � Viviparidae �Biogeography � Cipangopaludina �Bellamya japonica

Introduction

The ecological impacts of an invasive species are a

function of its range, abundance, and per-capita

effects, and may be manifested at levels of biological

Electronic supplementary material The online version ofthis article (doi:10.1007/s10530-009-9572-7) containssupplementary material, which is available to authorized users.

C. T. Solomon (&) � M. J. Vander Zanden

Center for Limnology, University of Wisconsin, 680 N.

Park St., Madison, WI 53706, USA

e-mail: ctsolomon@wisc.edu

J. D. Olden

School of Aquatic and Fishery Sciences, University

of Washington, Box 355020, Seattle, WA 98195, USA

P. T. J. Johnson

Department of Ecology and Evolutionary Biology,

University of Colorado, Ramaley N122, Campus Box 334,

Boulder, CO 80309, USA

R. T. Dillon Jr.

Department of Biology, College of Charleston,

Charleston, SC 29424, USA

123

Biol Invasions (2010) 12:1591–1605

DOI 10.1007/s10530-009-9572-7

organization ranging from the genome to the ecosys-

tem (Parker et al. 1999). For researchers and man-

agers interested in conserving species diversity,

understanding impacts at the community level is

particularly important. An emerging consensus sug-

gests that the risk of an invasive species causing

ecological change depends on the degree of func-

tional similarity between that species and the native

species pool; distinctiveness of the invader enhances

its impact on native species (Ricciardi and Atkinson

2004), and competitors rarely cause extinctions,

whereas predators are more likely to do so (Davis

2003; Sax et al. 2007). However, there remains a high

level of taxonomic and geographic bias in the study

of invasive species (Pysek et al. 2008), and relatively

few studies have quantified the community-level

consequences of invasive animals (Parker et al.

1999).

Freshwater gastropod assemblages provide a use-

ful and important model system for studying the

community-level effects of invasive competitors.

Field studies of gastropods show patterns consistent

with interspecific competition, and experimental

manipulations have demonstrated the importance of

interspecific resource competition and other negative

interactions within this group (Brown 1982; Hershey

1990; Adam and Lewis 1992; Turner et al. 2007; see

review in Dillon 2000). Furthermore, freshwater

gastropods are of significant conservation importance

because they are both highly diverse and highly

threatened (Lydeard et al. 2004; Strayer 2006; Lysne

et al. 2008). Concurrently, many gastropods have

become invasive; for instance, the United States

Geological Survey’s Nonindigenous Aquatic Species

database lists *40 invasive gastropod species in the

United States (http://nas.er.usgs.gov). Yet while some

invasive gastropods are known to have large effects

on invaded ecosystems (e.g., Hall et al. 2003;

Carlsson et al. 2004) or on particular native species

(e.g., Pointier and McCullough 1989), their commu-

nity-level effects on potential competitors within the

snail assemblage remain relatively unexplored.

The Chinese mystery snail (Bellamya [=Cipan-

gopaludina] chinensis (Reeve 1863)) is a viviparid

gastropod native to Asia. It is very large, reaching a

shell length of *65 mm and dry tissue mass of *1 g

(Jokinen 1982; Solomon unpubl. data). Its first

reported occurrence in North America was in a food

market in San Francisco in the early 1890s (Wood

1892). Since that time, it has spread across much of

the United States and parts of southern Canada, with

occurrences in at least 27 states plus Quebec (Jokinen

1982, http://nas.er.usgs.gov). Limited published lit-

erature, anecdotal accounts, and our own observa-

tions suggest that it can reach extremely high

densities in systems where it occurs. Due to its wide

distribution and high densities, some authors have

speculated that it might have significant impacts in

invaded systems (Bury et al. 2007), and experimental

studies indicate it can negatively affect native gas-

tropods (Johnson et al. 2009). However, very little is

known about the distribution of B. chinensis within

invaded regions, and no study has attempted to

quantify its impacts in invaded ecosystems.

In this study, we surveyed snail assemblages in 44

lakes in a region known to be invaded by B. chinensis.

We sought to describe the patterns and determinants of

B. chinensis distributions, and to assess whether

invasions of B. chinensis have altered the composition

of native snail assemblages. Because impacts of

invasive species may be localized or widespread, and

may occur rapidly or gradually, we considered multi-

ple spatial and temporal scales in our analysis. We

tested for impacts of B. chinensis on native snails at

both the lake and site (within lake) scale, and combined

our own data with historical survey data from the 1930s

to test whether contemporary assemblage structure or

long-term changes in assemblages were related to the

occurrence or abundance of B. chinensis.

Methods

Study system

We surveyed snail assemblages in lakes of the

Northern Highlands Lake District (NHLD), Wiscon-

sin, USA. There are [7,500 lakes in this region,

which vary widely with regard to morphology,

chemistry, human use, and other factors (Kratz

et al. 1997; Hanson et al. 2007). These lakes are an

important cultural and economic resource, and inva-

sive species are a primary management concern.

Recent reports indicate that B. chinensis is widely

distributed across Wisconsin and Minnesota (Jass

2004; Bury et al. 2007; Johnson et al. 2009), and

anecdotal accounts and our own observations indicate

that it has been present in some lakes in the NHLD

1592 C. T. Solomon et al.

123

since at least 2003. However, the date of first

occurrence of B. chinensis in the NHLD can be

given only generally based on current knowledge, as

some time between 1940 and 2000. The earliest

records of the species in the Great Lakes and Upper

Mississippi region are from the late 1930s and 1940s

(Mills et al. 1993; Bury et al. 2007), and the earliest

Wisconsin record is from the 1950s (Teskey 1954).

The taxonomy of introduced Bellamya in North

America is somewhat uncertain. Some authors have

recognized two species (B. chinensis and B. japon-

ica), whereas others have argued that the two forms

are simply variants of B. chinensis. For many years,

taxonomists placed the species in the genus Cipan-

gopaludina. We follow Smith (2000), who provided a

complete synonymy and the most recent taxonomic

revision, affirming the two-species concept and

placing the species in the genus Bellamya.

Field sampling

During the summer of 2006 we collected snails at 4–6

sites in each of 44 focal lakes. Survey lakes ranged in

surface area from 14 to 1,400 ha (median = 130 ha),

and were selected to span broad gradients of land-

scape position, water chemistry, human use, and other

characteristics, and to maximize overlap with lakes

where Morrison (1932) has previously described snail

assemblages. For each lake, site locations were chosen

randomly within each compass quadrant of the

shoreline, using Geographic Information System

(GIS) software (ArcGIS 9.2; ESRI, Redlands, Cali-

fornia). At each site we placed a 20 m transect line on

the lake bottom along the 1 m depth contour. At 2 m

intervals along the transect, two snorkelers collected

all the snails from within 0.25 m2 quadrats (10

quadrats per site). They then searched the vicinity of

the transect haphazardly for 5 minutes to reduce the

likelihood that B. chinensis presence at a site escaped

detection. Sampling ceased after the fourth site if at

least 25 quadrats with non-zero snail abundances had

been sampled; otherwise, sampling continued at

alternate sites until that threshold was reached or 6

sites had been sampled. At some sites the entire 20-m

transect fell in thick macrophyte beds, precluding

effective snorkel surveys. In these cases we sampled

snails by vigorously sweeping a D-net (500 lm mesh)

through the macrophytes in two 1 m2 areas. Collected

snails were preserved in 80% ethanol. Identifications

were made according to Burch (1989), following the

revision of Hubendick (1951) for the Lymnaeidae,

Hubendick (1955) for the Planorbidae, and Wething-

ton and Lydeard (2007) for the Physidae. All samples

are being curated into the Illinois Natural History

Survey Mollusk Collection.

We also tested the effectiveness of a rapid

assessment protocol for detecting the presence of

B. chinensis. Two observers snorkeled around the

vicinity of the boat launch (if present) for up to 5 min

each, or until B. chinensis was found. We conducted

this rapid assessment at 27 of the focal survey lakes,

as well as at 8 additional lakes where we did not

conduct full quadrat surveys.

Statistical analyses

The raw data from our sampling consisted of

abundances (counts) for each species in each quadrat.

For most analyses we aggregated these data, as either

the summed abundance at the site level or the mean

density at the lake level. Unless otherwise noted, all

statistical analyses were conducted in the R statistical

package (R Development Core Team 2008); the

‘‘vegan’’ package was used for multivariate analy-

ses (Oksanen et al. 2008).

We assessed the adequacy of our sampling design

in two ways. First, we used linear regression to test

for a relationship between sampling effort and lake-

level species richness. Sampling effort varied from a

minimum of 8 quadrats per lake (in lakes where all

sites occurred in dense macrophytes, and only net

sweep samples were taken) to a maximum of 60

quadrats per lake (in lakes where snails were rare or

absent, and 10 quadrats were sampled at each of 6

non-vegetated sites). Second, we developed species-

effort curves by rarefaction. Within each lake, we

constructed 500 bootstrap samples of the quadrat-

level data at each possible level of sampling intensity

(from 1 quadrat up to the number actually sampled in

that lake). We plotted the mean bootstrapped species

richness against sampling effort for each lake.

We described the probability of B. chinensis occur-

rence in a lake using multiple logistic regression

(a generalized linear model with a logit link function).

We considered eight predictors describing lakes in

terms of their physical and chemical characteristics

(area, conductivity, and Secchi depth), potential

anthropogenic dispersal vectors (shoreline housing

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1593

123

density, distance to population center, and accessibility

to boats), and the presence of other invasive species

that might interact with B. chinensis (rusty crayfish

Orconectes rusticus and banded mystery snail Vivip-

arus georgianus). Sources and methods for these data

are described in the Appendix (Table A1—Electronic

Supplementary Material). Continuous predictor vari-

ables were log-transformed when necessary to nor-

malize distributions, and all continuous predictors

were transformed to Z-scores. We used all-subsets

regression based on Akaike’s Information Criterion to

identify the best subsets of predictors, assessed the

significance of each term in the selected model using

likelihood ratio tests, and described the predictive

power of the model by quantifying the area under the

receiver operating characteristic curve (ROC AUC) of

sensitivity versus 1-specificity (Agresti 2002).

We also considered whether within-lake distribu-

tions of B. chinensis were related to the relative

proximity of boat launches. For all lakes where

B. chinensis was present, we calculated the shoreline

distance between each survey site and the nearest boat

launch using a Geographic Information System (GIS).

To allow comparisons between lakes of different sizes

that may have been invaded at different times, we

relativized all of the site-to-launch distances for each

lake by dividing by the maximum site-to-launch

distance for that lake. Thus all the distances for all

lakes ranged between 0 and 1. We then used logistic

regression to test the hypothesis that the likelihood of

B. chinensis presence at a site was higher for sites close

to boat launches. We excluded from this analysis one

lake where B. chinensis was present at every site.

Potential effects of B. chinensis on the native snail

assemblage were examined using multivariate analy-

ses. First, we used non-metric multidimensional scal-

ing (NMDS) to reduce the dimensionality of the

community data and display how dominant gradients

of variation in species composition were related to

environmental variables and to the presence and

abundance of B. chinensis. NMDS is an ordination

method that preserves the ranked-ordered distances

between sample points in ordination space by mini-

mizing a measure of disagreement (referred to as

stress) between the compositional dissimilarities and

the distance between points in the ordination diagram

(Kruskal 1964). In the context of our study, we were

particularly interested in separating the effects on

assemblage structure due to B. chinensis from those

due to environmental variables and other invasive

species. In our study lakes, conductivity is a primary

environmental determinant of snail assemblage struc-

ture (see Results), and the rusty crayfish (Orconectes

rusticus) and the banded mystery snail (Viviparus

georgianus) are invasive species that might also

influence snail assemblages. We therefore tested for

an effect of B. chinensis on assemblage structure while

controlling for conductivity and the presence of O.

rusticus and V. georgianus. This analysis was con-

ducted using permutation MANOVA (perMANOVA;

(Anderson 2001), which allows for ANOVA-style

partitioning of variance among predictors when the

response is a multivariate community distance matrix.

Both NMDS and perMANOVA consider the ranked

dissimilarities in assemblage structure among loca-

tions, which we calculated using Sørensen’s (Bray-

Curtis) distance metric. Prior to the analyses, species

abundance data were log(x ? 1) transformed and

relativized by species maxima. NMDS ordinations

were conducted in the PC-ORD software package

(McCune and Mefford 1999), and perMANOVA tests

were performed using the ‘‘adonis’’ function in R. We

also inspected plots of the abundance of each species

versus the abundance of B. chinensis for evidence that

higher B. chinensis abundances were associated with

lower abundances of native species.

We quantified changes in lake-level snail assem-

blages from 1930 (prior to invasion by B. chinensis) to

the present and considered whether these changes

were related to invasion by B. chinensis. Historical

snail assemblage data were available for 30 of our

survey lakes (Morrison 1932). To make those data

comparable with our own, two differences in methods

and objectives had to be reconciled. First, advances in

freshwater gastropod systematics over the last

75 years dictated a substantial simplification of Mor-

rison’s taxonomy. Second, Morrison’s paper focused

on compiling a list of the localities where species

where known to occur, rather than describing the snail

assemblages in lakes. The omission of a lake from his

list of localities for a species could therefore represent

merely a lack of data, rather than a true absence of that

species in that lake. To account for this we classified

snail occurrences (species-lake combinations)

recorded by Morrison as either retained (still present

in our survey of that lake) or lost (absent from our

survey of that lake), and ignored occurrences in our

own data that were not recorded by Morrison. We

1594 C. T. Solomon et al.

123

used logistic regression to test whether the likelihood

that occurrences were lost was predicted by the

current-day presence of B. chinensis. In this analysis

we controlled for other factors that might influence

the likelihood that an occurrence was lost, including

the presence of the invasive snail predator O. rusticus;

the presence of the invasive snail V. georgianus; and

contemporary shoreline housing density, an indicator

of the extent of human alteration of littoral zones,

which is the primary anthropogenic impact on these

lakes (Carpenter et al. 2007). The Morrison data and

the species synonomy that we used are available in

Tables A2 and A3 (Electronic Supplementary

Material).

Results

Sampling efficiency and snail community

composition

We collected and identified 17,516 snails representing

21 species (Table A4—Electronic Supplementary

Material). Our sampling strategy effectively repre-

sented the species membership of the snail assem-

blages. There was no evidence that we found more

species when effort was higher; in fact, there was a

negative relationship between species richness and

the number of quadrats sampled (F1,42 = 25.4,

P \0.0001, R2 = 0.38), suggesting that the adaptive

sampling strategy was effective in concentrating the

greatest effort in lakes where snails were most difficult

to find. Rarefaction analysis demonstrated that sam-

pling effort was sufficient to find most or all of the

species present in the habitats that we surveyed. For

every lake, species-effort curves reached an asymptote

at or near the species richness that we observed in our

samples, often at a sampling intensity less than the

number of quadrats that we actually surveyed in the

lake (Fig. A1—Electronic Supplementary Material).

Species abundances were approximately log-nor-

mally distributed, with a few very abundant species

and many rare species. For instance, the two most

numerically-dominant species (Amnicola limosa and

Marstonia lustrica) comprised 66% of total abun-

dance. Lake-level species richness ranged from 0

(Camp Lake and Crystal Lake) to 14 (Allequash Lake

and Tomahawk Lake). The best model describing

species richness, selected by AIC using all-subsets

regression, included Secchi depth and conductivity

but not log(area), B. chinensis presence, O. rusticus

presence maximum depth, or an indicator for whether

only vegetated sites were sampled in the lake

(richness = -0.75 9 Secchi ? 0.05 9 conductivity ?

6.0; P\0.0001, R2 = 0.45).

Distribution of B. chinensis within

and among lakes

B. chinensis was widely distributed across the study

region (Fig. 1). Of the 42 focal lakes in which snails

were present, we observed B. chinensis in quadrat

samples in 15 lakes, and during haphazard searching

near quadrats in an additional 6 lakes. We also

observed B. chinensis in 2 of the 8 lakes where only

the rapid assessment protocol was conducted. The

rapid assessment protocol was fairly effective, iden-

tifying B. chinensis presences at 9 of 12 lakes where

it was used in conjunction with regular quadrat

sampling and where B. chinensis was detected by

either method.

There was a strong geographic pattern in

B. chinensis occurrences, with the likelihood that a

lake was invaded much higher in the south than in the

north of the study region (Fig. 1). Invaded lakes in

our survey occurred in the Chippewa River and

Wisconsin River portions of the Mississippi River

drainage basin and not in the Lake Superior basin.

However, we have observed B. chinensis in lakes in

the Superior basin that were not among the 3 lakes in

that basin that we surveyed here (Johnson et al.

2009). The best model (AIC = 59.9) describing B.

chinensis presence indicated that it was more likely to

occur in lakes that were closer to a population center,

or that had higher shoreline housing density or lower

water clarity (Table 1). This model had good predic-

tive power (area under the receiver operating char-

acteristic curve = 0.83), illustrating a correct

classification rate of 78 % (based on a decision

threshold of 0.5). Good alternative models either

added a positive conductivity term to those listed

above (AIC = 60.1) or selected the conductivity term

in place of the Secchi depth term (AIC = 61.0).

When present, B. chinensis was patchily distrib-

uted within lakes. In all but one of the fifteen lakes

where we detected B. chinensis in quadrat sampling,

it occurred in between 2% and 56% of quadrats

(mean 15% ± 16% SD), and mean densities ranged

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1595

123

between 0.16 and 4.00 individuals m-2 (mean

0.81 ± 1.04 SD). Otter Lake was a clear exception

to this pattern; in this lake, B. chinensis was present at

all four surveyed sites, and in 15 of 16 (94%) of

quadrats. The mean density of B. chinensis in Otter

Lake was 38 individuals m-2. Multiplying these

densities by an estimate of the mean biomass of

B. chinensis (mean ± SD of dry tissue mass exclud-

ing shell was 0.27 ± 0.15 g for a random sample of

30 B. chinensis collected from Allequash Lake in

2004; Solomon unpubl. data) yields biomass

estimates of 10.33 g dry mass m-2 in Otter Lake

and 0.04–1.08 g dry mass m-2 in the other lakes.

Within invaded lakes, the probability that

B. chinensis occurred at a given site was higher for

sites closer to boat launches (logistic regression,

v21 ¼ 6:1, P = 0.01; Fig. 2a). There was also weak

evidence that B. chinensis was more likely to occur in

lakes with improved public boat launches than in

lakes where public access was absent or limited to

carry-in boats (Pearson’s v2 = 2.8, P = 0.09;

Fig. 2b).

Effects of B. chinensis on native snail

assemblages

High site-level densities of B. chinensis were asso-

ciated with low densities of some native species

(Fig. 3). Specifically, Lyogyrus granum, the Valvata

species, the Lymnaea species, Physa acuta, and

Helisoma trivolvis tended not to occur at sites where

B. chinensis abundance was greater than between 0

and 2 individuals m-2. The three species of Lymnaea

together occurred at 19 sites where B. chinensis was

absent, but at only 3 sites where B. chinensis was

present. In contrast, there was no evidence for a

negative relationship between the site-level abun-

dance of B. chinensis and that of its closest relatives

in these lakes, the viviparids Campeloma decisum

and Viviparus georgianus.

Fig. 1 Map of the Northern

Highlands Lake District,

showing lakes where the

Chinese mystery snail

Bellamya chinensis was

present (lakes filled in

black) and absent (lakes

outlined in black). Lakes in

grey were not surveyed

Table 1 Multiple logistic regression model describing the

probability of B. chinensis occurrence in a lake

Estimate SE P(v2)

(Intercept) -0.26 0.35

Secchi depth -0.84 0.40 0.02

Distance to Minocqua -0.71 0.37 0.04

Log (buildings km-1) 1.00 0.40 0.005

Variables were selected by AIC using an all-subsets procedure.

Estimate, point estimate for coefficient; SE, standard error of

estimate; P(v2), probability of statistical significance associated

with a likelihood ratio test of the hypothesis that the

coefficient = 0

Note: Candidate variables included those in the selected model

above, as well as log (lake area), conductivity, presence of a

public boat ramp, presence of rusty crayfish, and presence of

the invasive snail Viviparus georgianus. All continuous

predictors were Z-transformed

1596 C. T. Solomon et al.

123

Despite the negative association between the site-

level abundances of B. chinensis and some native snail

species, there was no evidence that B. chinensis

affected overall snail assemblage structure at the level

of either sites or lakes (Fig. 4). Total native species

richness was, if anything, higher at sites where

B. chinensis was present (mean = 6.0, n = 20) than

at sites where only native species were present

(mean = 4.8, n = 123; t23 = 1.7, P = 0.1). Assem-

blage structure at the lake level was correlated with

environmental variables, particularly Secchi depth and

conductivity (Fig. 4b). Results from the perMANOVA

analysis confirmed the patterns suggested by the

NMDS ordinations; even after controlling for conduc-

tivity and for the presence of O. rusticus and

V. georgianus, there was no effect of the presence of

B. chinensis on lake-level snail assemblages (Table 2).

Similarly, while there were significant changes in

snail assemblages between the historical surveys and

our own, there was no apparent effect of B. chinensis

invasion on these changes. Many of the snail

occurrences (species-lake combinations) that were

present in the historical data were ‘‘lost’’ (that is,

were not detected by our surveys; see Fig. 5, where

most points fall below the 1:1 line). The likelihood

that an historical occurrence was lost varied among

species and families (Table 3; Fig. 5a). Species that

were rare in the historical data (particularly lymnae-

ids, ancylids, and valvatids) often were not detected

in any of the lakes that they historically occupied

(Fig. 5a), although two of these (Valvata sincera and

Promenetus exacuous) were detected in lakes that

they had not been recorded in historically. The

likelihood that an historical occurrence was lost did

not vary as a function of present-day shoreline

housing density, of main effects of the presence of

B. chinensis, O. rusticus, or V. georgianus, or of

interaction terms of those main effects with species

(Table 3; Fig. 5b).

Discussion

Distribution of B. chinensis

The results of our regional survey and the findings of

Bury et al. (2007) for Minnesota demonstrate that

B. chinensis occurs frequently within invaded

regions, besides being widely distributed across the

United States (Jokinen 1982, http://nas.er.usgs.gov).

In the Northern Highlands Lake District of Wiscon-

sin, B. chinensis occurred in 50% of the 42 lakes in

which we detected any snails with our full sampling

protocol, making it among the most frequently

occurring species in our study. Furthermore, the

A

Relative distance to boat launch

Num

ber

of s

ites

0

4

8

12

16 B. chinensis absentB. chinensis present

B

0.0 0.2 0.4 0.6 0.8 1.0

with boat ramp without boat ramp

Num

ber

of la

kes

0

5

10

15

20

Fig. 2 Relationship between Bellamya chinensis occurrence

and public boat launches. a The probability of B. chinensisoccurrence at sites within lakes where it occurred was highest

for sites close to a boat launch (P = 0.01). Sites were binned

according to their relative distance to the nearest boat launch

(see text). Data are from 14 lakes where B. chinensis was

detected in quadrat sampling (excludes one lake where

B. chinensis was found at all sites). b B. chinensis was more

likely to occur in lakes with improved public boat launches

than in lakes without launches or where access was limited to

carry-in boats only (P = 0.09)

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1597

123

strong south-to-north pattern in occurrences and the

link between occurrences and boat launches suggest

that this species has not yet reached all of the lakes in

the NHLD in which it could potentially persist. While

our focal lakes were selected with some bias towards

lakes where previous studies had observed snails,

they were otherwise broadly representative of lakes in

the region. It may therefore be reasonable to expect

that 50% or more of the accessible lakes in this region

are or could be invaded by B. chinensis. In compar-

ison, maximum potential distributions of many

aquatic invasive species seem to be near or below

half of the lakes in a region. For instance, 41% of 179

Minnesota lakes were estimated to be susceptible to

invasion by spiny water flea Bythotrephes longim-

anus (Branstrator et al. 2006), 55% of 8,110 Ontario

lakes and 11% of 5164 Wisconsin lakes were esti-

mated to be susceptible to invasion by rainbow smelt

Osmerus mordax (Mercado-Silva et al. 2006), and

9.9% of 3908 Wisconsin lakes are deemed environ-

mentally suitable for rusty crayfish Orconectes rust-

icus (Olden, unpubl. data). In regions where it occurs,

B. chinensis may be among the most ubiquitous of

invasive aquatic animals.

An important research need for managing any

invasive species is to identify which systems are

susceptible to invasion. Our analyses indicated that

the likelihood of B. chinensis occurrence was influ-

enced both by intrinsic properties of lakes (Secchi

depth and conductivity) and by human activities

(distance to population center, shoreline housing

density, and boat launches). Apparent relationships

between environmental characteristics and the occur-

rence of B. chinensis should be interpreted with

caution given the suggestion above that this species

has not yet saturated the landscape in the NHLD

(Peterson 2003). With that caution in mind, we

discuss below what general properties of lakes

1

10

100

1000 Vivgeo

1

10

100

1000 Valtri

1

10

100

1000 Phyacu

1

10

100

1000 Proexa

Camdec Amnlim Amnlus Lyogra

Gyrpar Gyrdef Ferpar

1 10 100 1 10 100 1 10 100 1 10 100 1 10 100

Ferriv

Valsin Lymbul Lymcat Lymsta

Phygyr Helanc Helcam Heltri

B. chinensis abundance + 1

Spe

cies

abu

ndan

ce +

1

Fig. 3 Abundance of each native snail species (vertical axis)

and B. chinensis (horizontal axis) at 197 sites in 42 lakes. Note

log scale on axes. For each panel, the species plotted on the

y-axis is indicated in the top right corner. Species are Amnicolalimosa (Amnlim), Marstonia lustrica (Marlus), Lyogyrusgranum (Lyogra), Valvata sincera (Valsin), Valvata tricarinata(Valtri), Bellamya chinensis (Belchi), Campeloma decisum

(Camdec), Viviparus georgianus (Vivgeo), Ferrissia parallela(Ferpar), Ferissia rivularis (Ferriv), Lymnaea bulimoides(Lymbul), Lymnaea catascopium (Lymcat), Lymnaea stagnalis(Lymsta), Physa acuta (Phyacu), Physa gyrina (Phygyr),

Gyraulus deflectus (Gyrdef), Gyraulus parvus (Gyrpar),

Helisoma anceps (Helanc), Helisoma campanulata (Helcam),

Helisoma trivolvis (Heltri), and Promenetus exacuous (Proexa)

1598 C. T. Solomon et al.

123

ecosystems may increase their susceptibility to

B. chinensis invasion.

The relationships between B. chinensis occurrence

and lake conductivity and Secchi depth may indicate

an affinity for more productive systems, and/or

minimum Ca requirements for shell growth. Both

Secchi depth and conductivity are correlated with

lake productivity, and more productive systems may

allow greater snail species richness (and greater

likelihood of B. chinensis occurrence) because of

greater food availability (Dillon 2000). Conductivity

is also strongly correlated with Ca concentration in

these lakes (e.g., Johnson et al. 2008). Calcium

requirements for B. chinensis are likely to be high due

to its large shell, and Ca limitation has been

suggested to be a primary determinant of occurrence

for other invasive mollusks (e.g., Whittier et al.

2008). A survey of [200 Connecticut water bodies

found B. chinensis only in those with Ca concentra-

tions [5 mg l-1, even though 75% of the waters

surveyed had Ca concentrations lower than that

threshold (Jokinen 1982). We found B. chinensis in

some systems that likely had Ca concentrations

\5 mg l-1 (based on the relationship between con-

ductivity and Ca in 50 NHLD lakes; data not shown),

but nonetheless the positive relationship between

conductivity and the likelihood of B.chinensis occur-

rence could indicate a Ca limitation effect. It is also

important to note that conductivity is correlated with

many important properties of lakes in this region

(Kratz et al. 1997), and so its relationship with

B. chinensis occurrence could be masking the effects

of other factors. For instance, lakes with high

conductivity may also be those most frequently

visited by boaters, since they tend to be large and

support recreational fishing opportunities (Reed-

Andersen et al. 2000).

The relationships between B. chinensis occurrence

and distance to population center, shoreline housing

density, and boat launches indicate that greater

human use of a lake is associated with a higher

Axis 1 (25.9% of variation)

Axi

s 2

(22.

3% o

f var

iatio

n)A

Axis 1 (41.7% of variation)

Axi

s 2

(21.

5% o

f var

iatio

n)

secchi cond

B

Fig. 4 Non-metric multidimensional scaling (NMDS) of snail

assemblage data, showing the two axes (of three total) that

explained the most variation in assemblage structure in each

ordination. a Points represent assemblage structure at the site

level (143 sites in 42 lakes) for sites where B. chinensis was

absent (empty circles) or present (filed circles; diameter of

circle proportional to log(abundance) of B. chinensis at the

site). Final stress = 0.20. b Points represent assemblage

structure at the lake level (42 lakes). Symbols as in (a).

Radiating lines indicate the direction and strength of correla-

tions between ordination axis and environmental variables that

explained[10% of the variation along any axis (secchi, Secchi

depth, and cond, conductivity). Ordination has been rotated to

maximize the correlation between conductivity and Axis 1.

Final stress = 0.14

Table 2 perMANOVA tests of hypothesized factors influ-

encing snail assemblage structure, using lake-level data

df SS F P

Conductivity 1 0.91 3.06 \0.001

O. rusticus 1 0.36 1.21 0.24

V. georgianus 1 0.22 0.75 0.76

B. chinensis 1 0.29 0.97 0.47

Residual 37 10.96

Interaction terms are excluded for simplicity; none were

significant (all P [ 0.4). df, degrees of freedom; SS, sums of

squares; F, pseudo-F statistic for the perMANOVA test; P,probability of statistical significance (based on 1,000 permutations

of the data) associated with a test of the hypothesis that there is no

effect of a predictor. Terms for species are presence-absence of

three invasive species: Orconectes rusticus (rusty crayfish),

Viviparus georgianus (banded mystery snail), and Bellamyachinensis (Chinese mystery snail)

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1599

123

likelihood of B. chinensis establishment. The link

between boat launches and within-lake distributions

of B. chinensis suggests that boater movements in

particular play an important role in dispersal. Boats

are an important vector of spread for many aquatic

invasive species, and previous studies have observed

correlations between patterns of boater movement

and lake-level occurrences of invasive species (e.g.,

Buchan and Padilla 1999; MacIsaac et al. 2004).

B. chinensis individuals will attach to macrophytes

that could get tangled on boat trailers, and we have

observed individuals brought into boats inadvertently

with sediments on anchors. The ability of B. chinensis

to close its operculum probably makes it fairly

resistant to desiccation once on a boat or trailer; that

trait, and the fact that it bears live young which may

be ‘‘stored’’ for long periods inside the adult, could

facilitate invasions even when boaters do not visit

new lakes on the same day. Intentional releases from

aquaria and water gardens and dispersal along stream

corridors could also spread B. chinensis to new lakes.

We are not aware of previous studies that have

linked within-lake distributions of invasive species to

boat launches, as our results do for B. chinensis. Yet

this pattern may be common for animal and plant

species with limited mobility, as spread around a lake

may take years after establishment at an invasion point

(Wilson et al. 2004). If rates of spread were available,

this information would allow dates of invasion to be

back-calculated in cases where species had not yet

established throughout the lake. Furthermore, the

A

0 10 20

Cur

rent

occ

urre

nces

0

10

20

HydrobiidaeValvatidaeViviparidaeAncylidaeLymnaeidaePhysidaePlanorbidae

B

Historical occurrences100

0

10

B.c. 0, O.r. 0B.c. 0, O.r. 1B.c. 1, O.r. 0B.c. 1, O.r. 1

Fig. 5 Our surveys detected some but not all of the snail

occurrences (species-lake combinations) recorded by Morrison

(1932). a Number of lakes in which a species was found in our

surveys plotted against number of lakes in which it historically

occurred. Symbols are coded to indicate the family to which

each species belongs, including both prosobranchs (solidpoints) and pulmonates (hollow points). b Number of species

found in a lake in our surveys plotted against number of species

historically found there. Symbols are coded according to the

presence (1) or absence (0) of the invasive species Bellamyachinensis and Orconectes rusticus. Size of symbols is

proportional to present-day shoreline housing density at each

lake. Solid line indicates 1:1 relationship in each panel

Table 3 Results from a multiple logistic regression describing

the likelihood that species occurrences (species-lake combi-

nations) recorded by Morrison (1932) were lost (i.e., were not

detected in our surveys)

df Deviance Likelihood

ratio

P (v2)

Shoreline housing

density

1 148.2 1.66 0.2

Species 24 277.2 83.73 \0.0001

O. rusticus 1 193.5 0.001 1.0

V. georgianus 1 193.5 0.013 0.9

B. chinensis 1 193.6 0.067 0.8

Species 9 O. rusticus 17 165.6 19.0 0.3

Species 9 V. georgianus 16 162.6 16.1 0.4

Species 9 B. chinensis 19 163.4 16.8 0.6

The full model was focused on identifying effects of the

invasive Chinese mystery snail Bellamya chinensis on the

likelihood of loss. We controlled for among-lake differences in

human alterations to littoral zone habitat (as indicated by log-

transformed shoreline housing density); for differences among

native snail species; for possible effects of the invasive crayfish

Orconectes rusticus and the invasive snail Viviparusgeorgianus; and for the possibility that the effects of the

invasives on the natives varied among natives. The table gives

the degrees of freedom (df), deviance, likelihood ratio, and

probability of statistical significance (P) for likelihood ratio

tests of the significance of each term in the model

1600 C. T. Solomon et al.

123

strong link between boat launches and localized

distributions of B. chinensis raises an interesting

hypothesis. If boater movements play such an impor-

tant role in determining the distribution of B. chinensis,

it seems reasonable to expect that they could also affect

the distribution of native snails. It might be interesting

to ask whether entire snail assemblages in these lakes

are to some extent structured by patterns of human

movement across the landscape.

B. chinensis impacts on native snail assemblages?

Evidence that B. chinensis invasions influenced

native snail assemblages varied with the spatial scale

of analysis. At the whole-lake scale, neither the

contemporary data nor the historical comparison

showed any effects of B. chinensis on the presence

or abundance of native snails. These results are

consistent with the idea that invasive species rarely

extirpate their competitors (Davis 2003; Sax et al.

2007). At the more local scale of sites, the abundance

of B. chinensis was negatively associated with that of

several native species. This pattern might result from

competitive effects, but might simply reflect differ-

ences in niche requirements. Teasing apart these

alternative explanations is a notoriously difficult

problem in community ecology, although controlled

experiments can sometimes shed light on patterns

observed in the field. A recent mesocosm experiment

demonstrated that B. chinensis at densities *10

individuals m-2 had stronger effects on Lymnaea

stagnalis (reduced survival) than on Physa gyrina

(reduced growth) (Johnson et al. 2009). In our

surveys, L. stagnalis (but not P. gyrina) was one of

the species that did not occur at high abundance at

sites where B. chinensis was present (Fig. 3). This

qualitative agreement between experimental and

observational results lends some support to the idea

that B. chinensis has had negative effects on at least

some of the species with which it is negatively

associated. On the other hand, any such effects are

relatively subtle, in that no B. chinensis effect on

overall assemblage structure at the site level was

detectable (Fig. 4b). On balance our results provide,

at most, weak evidence that B. chinensis negatively

impacts native snails.

Why do we not see strong evidence that B. chinensis

invasion has negative effects on native snail popula-

tions? At least three explanations are plausible. First, it

could be that insufficient time has elapsed since

invasion to detect the effects of what are in fact strong

competitive interactions (Strayer et al. 2006). This

explanation seems fairly unlikely given that no

B. chinensis effects were observed in the re-survey

analysis. However, the temporal perspective gained

from that analysis goes back only as far as the invasion

of a given lake, and it is possible that some of our

survey lakes were invaded relatively recently. Fur-

thermore, the re-survey analysis can only detect

B. chinensis effects if they result in lake-wide extir-

pation of a native, whereas significant impacts could

occur below that threshold. Indeed, a second explana-

tion for the lack of strong evidence for a negative

impact is that the impacts are not strong, but rather are

weak and localized.

Finally, of course, it could be that B. chinensis

simply does not compete with native snails. This

would suggest either resource partitioning among the

species or that resources are not limiting. Resource

partitioning seems unlikely; all of the snails collected

in our survey were co-occurring at fine spatial scales,

suggesting minimal potential for diet specialization

given that most snails seem to consume diet items

unselectively within the constraints imposed by

habitat selection (Dillon 2000). Some members of

the Viviparidae will filter feed from the water

column, but stable isotope ratios of B. chinensis

collected from one of our study lakes suggest heavy

reliance on benthic resources and little if any reliance

on pelagic resources (C. Solomon, unpubl. data). If

these species are indeed generalized deposit feeders,

then a quick glance at the lake bottom supports the

hypothesis that resources are not limiting: in most

locations in most lakes in the NHLD, an ample layer

of organic detritus mixed with periphyton covers the

bottom. This provides an interesting contrast with

some stream ecosystems where the invasive New

Zealand mud snail (Potamopyrgus antipodarum)

reaches biomasses up to 10–30 g ash-free dry

mass m-2 and appears to reduce production of native

invertebrates by appropriating much of the available

primary production (Kerans et al. 2005; Hall et al.

2006). Differences between streams and lakes in the

size of organic matter pools may help to explain this

apparent difference in the effects of B. chinensis and

P. antipodarum.

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1601

123

Structure of snail assemblages

What factors do control snail assemblage structure in

this region? Of those that we considered, conductivity

and Secchi depth were the most important (Fig. 4b).

Increased availability of food (via greater primary

production), increased availability of Ca for shell

growth, and increased connectivity to dispersal corri-

dors in drainage lakes with higher conductivity are all

possible explanations for these effects. A number of

studies have identified similar patterns, particularly

between species richness and these indicators of lake

ion content and productivity (Dillon 2000). Other

factors identified as important determinants of snail

assemblage structure in previous studies include

habitat diversity (Aho 1966; Harman 1972; Bronmark

1985), predators (Lodge et al. 1987), and dispersal

(Bronmark 1985; Lewis and Magnuson 2000; Heino

and Muotka 2006). We did not consider the effects of

habitat diversity, but we did consider the effects of at

least one important snail predator, the invasive rusty

crayfish Orconectes rusticus.

Correlative studies, small-scale experiments, and

multi-year monitoring of lakes invaded by rusty

crayfish have shown that they have strong effects on

the abundance of snails (e.g., Lodge et al. 1994; Lodge

et al. 1998; McCarthy et al. 2006). However, we found

no evidence that the presence of rusty crayfish

influenced contemporary snail assemblages or the

likelihood that historical species occurrences were lost

from lakes (Table 3). Similarly, a recent study com-

paring current snail abundances in lakes with and

without rusty crayfish to abundances measured in the

1930s found only weak evidence of an effect (Rosen-

thal et al. 2006). These apparently contradictory results

may indicate that transient dynamics imply stronger

impacts than are eventually manifested; alternatively,

they may indicate the importance of context (e.g., lake-

specific densities of rusty crayfish populations) in

determining the outcome of an invasion. Either way,

they illustrate the importance of a nuanced and a long-

term perspective for understanding the impacts, if any,

of non-native species (Strayer et al. 2006).

Many of the snail occurrences recorded by Morri-

son (1932) were not observed in our surveys.

Regardless of whether species were rare or common

in Morrison’s data, they occurred in fewer lakes in our

surveys than they did in the historical data (Fig. 5).

This is not attributable to insufficient sampling, as

rarefaction analyses demonstrated that we detected

most or all of the species present in the habitats that

we surveyed. However, it is possible that Morrison’s

data include species that rarely occur at 1 m depth,

where our surveys were conducted. For instance,

some of the species that were recorded by Morrison

but not in our surveys, such as Planorbula armigera

and some of the Lymnaea, are typical of small

stagnant water bodies (Baker 1928). As one simple

way to control for this potential bias, we repeated our

analysis of these data after excluding species that we

rarely observed in our surveys (i.e., those found in

\10% of the 30 lakes in the data set). The results (not

shown) were very similar to those obtained for the full

data set (Table 3); that is, even when we considered

only those species that we frequently detected in

lakes, there were no significant effects of species

invasions or shoreline development on the likelihood

that species occurrences were lost from these lakes. It

is possible that inconsistent identification of species

(not just changes in nomenclature, which we

accounted for) contributed to the apparent changes

in assemblages, although it is hard to imagine how this

could produce the patterns that we observed. Further

research is needed, perhaps employing paleoecolog-

ical techniques, to clarify the patterns and understand

the drivers of the significant changes in snail assem-

blages that seem to have occurred in this region.

Future research

Given the widespread distribution of B. chinensis

across the United States, the lack of information

about its potential ecological impacts to date is

striking. Our results suggest that B. chinensis at the

densities that we observed has few impacts on native

snail assemblages in the study lakes that we exam-

ined. However, our results also demonstrate that this

species may be very widespread within some lake

districts, and may occur at very high biomass in

invaded lakes. In comparison to our estimate that

B. chinensis biomass is 0.04–1.08 g dry mass m-2 in

most of the invaded lakes we surveyed, the total

biomass of benthic insects at a similar depth in

nearby Crampton Lake is *1.2 g dry mass m-2

(Babler et al. 2008). Furthermore, B. chinensis may

occur at higher densities than we observed here. Thus

this species may be a significant component of the

benthic community in many lakes where it occurs.

1602 C. T. Solomon et al.

123

Further study of the impacts of this species on lake

biota and processes is therefore probably warranted.

Experimental evidence indicates that B. chinensis

may alter algal biomass and nutrient cycling in the

benthic community (Johnson et al. 2009). The

importance of benthic processes for structure and

function in lake ecosystems (Schindler and Scheue-

rell 2002; Vadeboncoeur et al. 2002) suggest that

alterations of these processes by B. chinensis could

have far-reaching consequences for invaded lakes,

which might suggest further management action

targeted at this species. Alternatively, if B. chinensis

has no appreciable effect on these processes, it might

be considered to be a relatively benign invasive

species, allowing management dollars to be targeted

on invasive species that do have clear negative effects

(Vander Zanden and Olden 2008). Distinguishing

between these alternatives is an important priority for

further research on this species.

Acknowledgments We thank K. Langree and E. Vennie-

Volrath for conducting the surveys. T. Kratz and the Trout

Lake Station staff supported the field work in a number of

ways, and conducted the regional surveys from which we

derived some of our data. B. Fahey and S. Jones provided

helpful statistical suggestions, L. Kursel and J. Maxted assisted

with statistical and GIS analyses, and J. Jass provided

information about occurrence records. This work was funded

by an NSF Graduate Research Fellowship to C. Solomon, by

grants from the Department of Zoology and the Center for

Limnology at the University of Wisconsin (including a Carl A.

Bunde award, an Anna Grant Birge award, and a Juday

fellowship), by the Wisconsin Department of Natural

Resources, and by the North-Temperate Lake Long-Term

Ecological Research (NTL-LTER) Program.

References

Adam ME, Lewis JW (1992) The lack of coexistence between

Lymnaea peregra and Lymnaea auricularia (Gastropoda,

Pulmonata). J Molluscan Stud 58:227–231

Agresti A (2002) Categorical data analysis, 2nd edn. Wiley,

Hoboken

Aho J (1966) Ecological basis of the distribution of the littoral

freshwater molluscs in the vicinity of Tampere, South

Finland. Ann Zool Fenn 3:287–322

Anderson MJ (2001) A new method for non-parametric mul-

tivariate analysis of variance. Austral Ecol 26:32–46

Babler AL, Solomon CT, Schilke P (2008) Depth-specific

patterns of benthic secondary production in an oligo-

trophic lake. J North Am Benthol Soc 27:108–119

Baker FC (1928) The fresh water Mollusca of Wisconsin. State

of Wisconsin, Madison

Branstrator DK, Brown ME, Shannon LJ, Thabes M, Heimgartner

K (2006) Range expansion of Bythotrephes longimanus in

North America: evaluating habitat characteristics in the

spread of an exotic zooplankter. Biol Invasions 8:1367–1379

Bronmark C (1985) Fresh water snail diversity—effects of

pond area, habitat heterogeneity and isolation. Oecologia

67:127–131

Brown KM (1982) Resource overlap and competition in pond

snails—an experimental analysis. Ecology 63:412–422

Buchan LAJ, Padilla DK (1999) Estimating the probability of

long-distance overland dispersal of invading aquatic spe-

cies. Ecol Appl 9:254–265

Burch JB (1989) North American freshwater snails. Malaco-

logical Publications, Hamburg

Bury JA, Sietman BE, Karns BN (2007) Distribution of the

non-native Viviparid snails, Bellamya chinensis and Vi-viparus georgianus, in Minnesota and the first record of

Bellamya japonica from Wisconsin. J Freshw Ecol 22:

697–703

Carlsson NOL, Bronmark C, Hansson LA (2004) Invading

herbivory: the golden apple snail alters ecosystem func-

tioning in Asian wetlands. Ecology 85:1575–1580

Carpenter SR, Benson BJ, Biggs R, Chipman JW, Foley JA,

Golding SA, Hammer RB, Hanson PC, Johnson PTJ,

Kamarainen AM, Kratz TK, Lathrop RC, McMahon KD,

Provencher B, Rusak JA, Solomon CT, Stanley EH,

Turner MG, Vander Zanden MJ, Wu C-H, Yuan H (2007)

Understanding regional change: a comparison of two lake

districts. Bioscience 57:323–335

Davis MA (2003) Biotic globalization: does competition from

introduced species threaten biodiversity? Bioscience 53:

481–489

Dillon RT Jr (2000) The ecology of freshwater molluscs.

Cambridge University Press, Cambridge

Hall R, Tank JL, Dybdahl MF (2003) Exotic snails dominate

nitrogen and carbon cycling in a highly productive stream.

Front Ecol Environ 1:407–411

Hall RO, Dybdahl MF, VanderLoop MC (2006) Extremely

high secondary production of introduced snails in rivers.

Ecol Appl 16:1121–1131

Hanson PC, Carpenter SR, Cardille JA, Coe MT, Winslow LA

(2007) Small lakes dominate a random sample of regional

lake characteristics. Freshw Biol 52:814–822

Harman WN (1972) Benthic substrates—their effect on

freshwater mollusca. Ecology 53:271–277

Heino J, Muotka T (2006) Landscape position, local environ-

mental factors, and the structure of molluscan assem-

blages of lakes. Landsc Ecol 21:499–507

Hershey AE (1990) Snail populations in arctic lakes—com-

petition mediated by predation. Oecologia 82:26–32

Hubendick B (1951) Recent Lymnaeidae. Their variation,

morphology, taxonomy, nomenclature, and distribution.

Kungl. Svenska Vetenskapskademiens Handlingar 3:1–

233

Hubendick B (1955) Phylogeny in the Planorbidae. Trans Zool

Soc Lond 28:453–542

Jass JP (2004) Distributions of gastropods in Wisconsin. Mil-

waukee Public Museum Contrib Biol Geol 99:1–28

Johnson PTJ, Olden JD, Vander Zanden MJ (2008) Dam

invaders: impoundments facilitate biological invasions

into freshwaters. Front Ecol Environ 6:359–365

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1603

123

Johnson PTJ, Olden JD, Solomon CT, Vander Zanden MJ (2009)

Interactions among invaders: community and ecosystem

effects of multiple invasive species in an experimental

aquatic system. Oecologia 159:161–170

Jokinen EH (1982) Cipangopaludina chinensis (Gastropoda:

Viviparidae) in North America, review and update. Nau-

tilus 96:89–95

Kerans BL, Dybdahl ME, Gangloff MM, Jannot JE (2005)

Potamopyrgus antipodarum: distribution, density, and

effects on native macroinvertebrate assemblages in the

greater Yellowstone ecosystem. J North Am Benthol Soc

24:123–138

Kratz TK, Webster KE, Bowser CJ, Magnuson JJ, Benson BJ

(1997) The influence of landscape position on lakes in

northern Wisconsin. Freshw Biol 37:209–217

Kruskal JB (1964) Multidimensional scaling by optimizing

goodness of fit to a nonmetric hypothesis. Psychometrika

29:1–27

Lewis DB, Magnuson JJ (2000) Landscape spatial patterns in

freshwater snail assemblages across Northern Highland

catchments. Freshw Biol 43:409–420

Lodge DM, Brown KM, Klosiewski SP, Stein RA, Covich AP,

Leathers BK, Bronmark C (1987) Distribution of fresh

water snails—spatial scale and the relative importance of

physicochemical and biotic factors. Am Malacol Bull 5:

73–84

Lodge DM, Kershner MW, Aloi JE, Covich AP (1994) Effects

of an omnivorous crayfish (Orconectes rusticus) on a

fresh water littoral food web. Ecology 75:1265–1281

Lodge DM, Stein RA, Brown KM, Covich AP, Bronmark C,

Garvey JE, Klosiewski SP (1998) Predicting impact of

freshwater exotic species on native biodiversity: chal-

lenges in spatial scaling. Aust J Ecol 23:53–67

Lydeard C, Cowie RH, Ponder WF, Bogan AE, Bouchet P,

Clark SA, Cummings KS, Frest TJ, Gargominy O, Herbert

DG, Hershler R, Perez KE, Roth B, Seddon M, Strong EE,

Thompson FG (2004) The global decline of nonmarine

mollusks. Bioscience 54:321–330

Lysne SJ, Perez KE, Brown KM, Minton RL, Sides JD (2008)

A review of freshwater gastropod conservation: chal-

lenges and opportunities. J North Am Benthol Soc 27:

463–470

MacIsaac HJ, Borbely JVM, Muirhead JR, Graniero PA (2004)

Backcasting and forecasting biological invasions of inland

lakes. Ecol Appl 14:773–783

McCarthy JM, Hein CL, Olden JD, Vander Zanden MJ (2006)

Coupling long-term studies with meta-analysis to inves-

tigate impacts of non-native crayfish on zoobenthic

communities. Freshw Biol 51:224–235

McCune B, Mefford MJ (1999) PC-ORD: multivariate analysis

of ecological data. MjM Software, Gleneden Beach

Mercado-Silva N, Olden JD, Maxted JT, Hrabik TR, Vander

Zanden MJ (2006) Forecasting the spread of invasive

rainbow smelt in the Laurentian Great Lakes region of

North America. Conserv Biol 20:1740–1749

Mills EL, Leach JH, Carlton JT, Secor CL (1993) Exotic

species in the Great Lakes—a history of biotic crises and

anthropogenic introductions. J Great Lakes Res 19:1–54

Morrison JPE (1932) A report on the Mollusca of the north-

eastern Wisconsin lake district. Trans Wisconsin Acad Sci

Arts Lett 27:359–396

Oksanen J, Kindt R, Legendre P, O’Hara B, Simpson G, Henry M,

Stevens H, Wagner H (2008) Community ecology package

(vegan). R Foundation for Statistical Computing, Vienna

Parker IM, Simberloff D, Lonsdale WM, Goodell K, Wonham

M, Kareiva PM, Williamson MH, Von Holle B, Moyle

PB, Byers JE, Goldwasser L (1999) Impact: toward a

framework for understanding the ecological effects of

invaders. Biol Invasions 1:3–19

Peterson AT (2003) Predicting the geography of species’

invasions via ecological niche modeling. Q Rev Biol 78:

419–433

Pointier JP, McCullough F (1989) Biological control of the

snail hosts of Schistosoma mansoni in the Caribbean area

using Thiara spp. Acta Trop 46:147–155

Pysek P, Richardson DM, Pergl J, Jarosik V, Sixtova Z, Weber

E (2008) Geographical and taxonomic biases in invasion

ecology. Trends Ecol Evol 23:237–244

R Development Core Team (2008) R: a language and envi-

ronment for statistical computing. R Foundation for Sta-

tistical Computing, Vienna

Reed-Andersen T, Bennett EM, Jorgensen BS, Lauster G,

Lewis DB, Nowacek D, Riera JL, Sanderson BL, Stedman

R (2000) Distribution of recreational boating across lakes:

do landscape variables affect recreational use? Freshw

Biol 43:439–448

Ricciardi A, Atkinson SK (2004) Distinctiveness magnifies the

impact of biological invaders in aquatic ecosystems. Ecol

Lett 7:781–784

Rosenthal SK, Stevens SS, Lodge DM (2006) Whole-lake effects

of invasive crayfish (Orconectes spp.) and the potential for

restoration. Can J Fish Aquat Sci 63:1276–1285

Sax DF, Stachowicz JJ, Brown JH, Bruno JF, Dawson MN,

Gaines SD, Grosberg RK, Hastings A, Holt RD, Mayfield

MM, O’Connor MI, Rice WR (2007) Ecological and

evolutionary insights from species invasions. Trends Ecol

Evol 22:465–471

Schindler DE, Scheuerell MD (2002) Habitat coupling in lake

ecosystems. Oikos 98:177–189

Smith DG (2000) Notes on the taxonomy of introduced Bell-amya (Gastropoda: Viviparidae) species in northeastern

North America. Nautilus 114:31–37

Strayer DL (2006) Challenges for freshwater invertebrate

conservation. J North Am Benthol Soc 25:271–287

Strayer DL, Eviner VT, Jeschke JM, Pace ML (2006) Under-

standing the long-term effects of species invasions. Trends

Ecol Evol 21:645–651

Teskey MC (1954) The mollusks of Brown County, Wisconsin.

Nautilus 68:24–28

Turner AM, Turner RR, Ray SR (2007) Competition and in-

traguild egg predation among freshwater snails: re-

examining the mechanism of interspecific interactions.

Oikos 116:1895–1903

Vadeboncoeur Y, Vander Zanden MJ, Lodge DM (2002) Putting

the lake back together: reintegrating benthic pathways into

lake food web models. Bioscience 152:44–54

Vander Zanden MJ, Olden JD (2008) A management frame-

work for preventing the secondary spread of aquatic

invasive species. Can J Fish Aquat Sci 65:1512–1522

Wethington AR, Lydeard C (2007) A molecular phylogeny of

Physidae (Gastropoda: Basommatophora) based on mito-

chondrial DNA sequences. J Molluscan Stud 73:241–257

1604 C. T. Solomon et al.

123

Whittier TR, Ringold PL, Herlihy AT, Pierson SM (2008) A

calcium-based invasion risk assessment for zebra and

quagga mussels (Driessena spp). Front Ecol Environ 6:

180–184

Wilson KA, Magnuson JJ, Lodge DM, Hill AM, Kratz TK,

Perry WL, Willis TV (2004) A long-term rusty crayfish

(Orconectes rusticus) invasion: dispersal patterns and

community change in a north temperate lake. Can J Fish

Aquat Sci 61:2255–2266

Wood WM (1892) Paludina japonica Mart. for sale in the San

Francisco Chinese markets. Nautilus 5:114–115

Distribution and community-level effects of the Chinese mystery snail (Bellamya chinensis) 1605

123